Coffee Project

Coffee Consumption Group Project

1. Install The Libraries:

library(dplyr)
library(ggplot2)
library(ggrepel) 
library(tidyverse)
library(stringr)
library(plotly)

2. Install The Data:

coffee <- read.csv("../../../../data/coffee_survey.csv")

3. Explore the data columns:

names(coffee)

 [1] "X"                            "age"                         
 [3] "cups"                         "where_drink"                 
 [5] "purchase_other"               "favourite"                   
 [7] "favorite_specify"             "additions"                   
 [9] "additions_other"              "sweetener"                   
[11] "style"                        "strength"                    
[13] "roast_level"                  "caffeine"                    
[15] "expertise"                    "coffee_a_bitterness"         
[17] "coffee_a_acidity"             "coffee_a_personal_preference"
[19] "coffee_b_bitterness"          "coffee_b_acidity"            
[21] "coffee_b_personal_preference" "coffee_c_bitterness"         
[23] "coffee_c_acidity"             "coffee_c_personal_preference"
[25] "coffee_d_bitterness"          "coffee_d_acidity"            
[27] "coffee_d_personal_preference" "prefer_abc"                  
[29] "prefer_ad"                    "prefer_overall"              
[31] "wfh"                          "total_spend"                 
[33] "know_source"                  "most_paid"                   
[35] "most_willing"                 "value_cafe"                  
[37] "spent_equipment"              "value_equipment"             
[39] "gender"                       "education_level"             
[41] "employment_status"            "number_children"             
[43] "political_affiliation"       

Now we can select some coffee-related columns to explore people’s preferences

Let’s start with favourite type of coffee:

unique(coffee$favourite)

 [1] "Pourover"                         "Cortado"                         
 [3] "Regular drip coffee"              "Iced coffee"                     
 [5] "Cappuccino"                       "Latte"                           
 [7] "Cold brew"                        "Americano"                       
 [9] "Espresso"                         NA                                
[11] "Other"                            "Mocha"                           
[13] "Blended drink (e.g. Frappuccino)"

Age is also a useful variable to explore

unique(coffee$age)

[1] "<18 years old"   ">65 years old"   "25-34 years old" "18-24 years old"
[5] "45-54 years old" "35-44 years old" "55-64 years old"

Unfortunately this data is poorly formatted…

4. Prepare the data

Clean up the favourites column

coffee <- coffee %>% 
    mutate(favourite = str_replace_all(favourite, c(
      "Blended drink \\(e.g. Frappuccino\\)" = "Blended", 
      "Regular drip coffee" = "Drip")))   

Clean up the age column

coffee <- coffee %>% 
    mutate(age = str_replace_all(age, c(
      " years old" = "", 
      ">65" = "65+")))  

Clean up the cups column

coffee <- coffee %>% 
    mutate(cups = str_replace_all(cups, c(
      "More than 4" = "5+", 
      "Less than 1" = "<1")))  

5. Visualise Coffee Preferences

First let’s see the distribution of favourite coffee types:

ggplot(data = coffee,
           mapping = aes(y = favourite)) + 
      geom_bar()

There’s still a lot going on here… We can simplify the data by grouping coffee types into broader preperation styles.

Group by Preperation Style

 PrepStyle <- coffee %>% 
      mutate(prep = case_when(
        favourite %in% c("Mocha", "Latte", "Cappuccino", "Cortado", "Espresso")
        ~ "Espresso",
        favourite %in% c("Iced coffee", "Cold brew", "Blended") 
        ~ "Iced",
        favourite %in% c("Drip", "Pourover", "Americano") 
        ~ "Brewed",
        .default = "other"))

Now we can visualise these groups

ggplot(data = PrepStyle,
           mapping = aes(x = prep)) + 
      geom_bar()

Reorder Preperation Styles

PrepStyle$prep <- factor(PrepStyle$prep, levels=c(
  'other', 'Iced', 'Brewed', 'Espresso')
  )

You can arrange the order of attributes to best visualise the data you are working with

Stacked Bar Graphs

Now we can add stacks to the bar graphs to explore relationships between preperation and other attributes

Preperation by Age

ggplot(data = PrepStyle,
           mapping = aes(x=age, fill=prep)) + 
      geom_bar(position='stack') +
      geom_text(aes(label = after_stat(count)), 
                stat = "count", position = "dodge", check_overlap = TRUE
      )

ggplot(data = PrepStyle,
      aes(fill=prep, x=age)) + 
  geom_bar(position='dodge') +
  geom_text(aes(label = after_stat(count)),
    stat = "count", position = position_dodge(.9)
  )

Preperation by Cups

ggplot(data = PrepStyle,
           mapping = aes(x=cups, fill=prep)) + 
      geom_bar(position='stack')

ggplot(data = PrepStyle,
       aes(fill=prep, x=cups)) + 
  geom_bar(position='dodge')

Preperation by Roast

ggplot(data = PrepStyle,
           mapping = aes(x=roast_level, fill=prep)) + 
      geom_bar(position='stack')

Preperation by Style

ggplot(data = PrepStyle,
           mapping = aes(y=style, fill=prep)) + 
      geom_bar(position='stack')